A Comparison of Empirical Performance : Pricing models vs pricing Kernels

Abstract

This study examines the empirical performance of the option pricing models and the pricing kernel-based models in the KOSPI 200 options market. We evaluate the pricing model, the pricing kernel-based GARCH option modle, Black-Scholes option pricing model, and the parametric pricing kernel models proposed by Rosenberg and Engle (2002) under the unified framework in which the underlying process follows the extended Duan (1995)-GARCH process. We find that empirical performance of the models is fairly improved when we estimate the parameters of models using the option data compared to the results when we only use the underlying data. The parametric pricing kernel of which functional form is a chebyshevpolynomial function shows the best pricing (forecasting) performance for all options among models we investigate. Unlike the pricing (forecasting) results, the difference of hedging performance between the models is relatively small.